Lowlevel temperature inversions

A prominent feature of the Arctic environment is the frequent occurrence of low-level temperature inversions (i.e., temperature increases with height). This was first demonstrated by Brooks (1931) from kite ascents over Siberia. More detailed studies from kite and captive balloon ascents made by Sverdrup (1933) during the Maud expedition provided some of the first detailed information on inversion structure. Wexler (1936) was the first to address physical controls behind the formation of Arctic inversions. Other early studies of inversion characteristics include Vowinkel and Orvig (1967, 1970), based on soundings from the NP stations (NP-4, NP-6 and NP-7) and Arctic coastal sites. More recent efforts include Kahl (1990), Overland and Guest (1991), and Serreze et al. (1992b).

The usual situation in the Arctic during winter is a surface-based inversion (typical for land) or an inversion above a shallow (30-80 m) mixed layer (typical for the ice covered ocean), a broad region of warm air with a temperature maximum of around 1000-1200 m, and a negative lapse rate aloft. During winter, inversions are evident in nearly all sounding profiles. The temperature difference between the inversion base and top averages 10-12 K. The frequency of inversions over land areas decreases through spring and summer, but they are still found in about 50% of soundings for June and July. Over the central Arctic Ocean, inversions are still present in the great majority of summer soundings. However, summer inversions are weaker than their winter counterparts, are thinner and tend to be elevated above the surface. Over the ocean, the inversion base is about 200-400 m above the surface during May-August, with the top located between 750 and 1000 m. The SHEBA observations showed an inversion persisting in summer at about 400 m altitude with an intensity of about 5 °C (Uttal et al., 2002). Over land, the depth of the summer mixed layer is greater.

Typical vertical temperature structures for winter are provided in Figure 5.12 based on averaged rawinsonde data from six island stations around the perimeter of the Arctic Ocean (Overland et al, 1997). Also illustrated (Figure 5.13) is the typical annual cycle of inversion characteristics based on over 6000 rawinsonde ascents for station Zhigansk

Figure 5.12 Mean temperature profiles for February 1987 from six stations located around the periphery of the Arctic Ocean:

(2) Chelyuskin (78° N,104° E), (3) Kotelny (76° N, 138° E), (4) Barrow (71° N, 86° W), (5) Mould Bay (76° N, 119° W) and (6) Eureka (80° N, 86° W) (from Overland et al., 1997, by permission of AMS).

Figure 5.12 Mean temperature profiles for February 1987 from six stations located around the periphery of the Arctic Ocean:

(2) Chelyuskin (78° N,104° E), (3) Kotelny (76° N, 138° E), (4) Barrow (71° N, 86° W), (5) Mould Bay (76° N, 119° W) and (6) Eureka (80° N, 86° W) (from Overland et al., 1997, by permission of AMS).

Jfmamjjasond Months

JFMAMJJASOND JFMAMJJASOMD

Month Month

Figure 5.13 Monthly median inversion top (top of bars), base (bottom of bars) and temperature difference (solid lines) from (a) drifting station data from the central Arctic Ocean; (b) station Zhigansk over the Siberian tundra (from Serreze et al., 1992b, by permission of AMS).

JFMAMJJASOND JFMAMJJASOMD

Month Month

Figure 5.13 Monthly median inversion top (top of bars), base (bottom of bars) and temperature difference (solid lines) from (a) drifting station data from the central Arctic Ocean; (b) station Zhigansk over the Siberian tundra (from Serreze et al., 1992b, by permission of AMS).

in Siberia (66.8° N, 123.4° E) and over 2000 ascents over the central Arctic Ocean from the NP program (Serreze etal., 1992b). The latter figure shows the predominance of strong surface or near-surface based inversions in winter and the weaker elevated inversions in summer. The shallow mixed layer typically present in winter over the Arctic Ocean is masked because of the limited resolution of the rawinsonde data very near the surface. This mixed layer is related to the conductive heat flux through the snow and ice. These are mean conditions - the inversion depth and intensity vary regionally and exhibit considerable variability on daily to weekly time scales.

The Arctic inversion has a number of important climate effects. An inversion represents strong vertical stability, limiting the depth of vertical mixing of sensible heat and moisture. As noted previously, condensate plumes emanating from wide open-water areas, that extend to 4 km in the atmosphere and persist for up to 200 km downwind, have been identified during winter. The frequency at which such events can occur is determined in part by the strength of the inversion layer. Elevated concentrations of pollution gases and aerosols have been observed to coincide with the top of the inversion layer (Bridgman etal., 1989). Photochemical destruction of boundary-layer ozone at Arctic sunrise (Barrie et al., 1988; Oltmans et al., 1989) appears to involve a process in which ozone-depleted air within the inversion layer is occasionally replaced by above-inversion, ozone-rich air, via mixing processes.

Wexler (1936) was the first to propose maintenance of the winter Arctic temperature inversion in terms of longwave radiative equilibrium. He considered an ideal sounding consisting of a surface inversion of infinitesimal thickness below an isothermal layer extending up to a nearly adiabatic temperature layer. Wexler hypothesized that the nearly blackbody emission of the snow surface was in radiative equilibrium with the partial emission of the isothermal layer. Put differently, because the atmosphere has a lower emissivity than the surface, radiative equilibrium requires that the surface be radiating at a lower physical temperature than the air in the isothermal layer. A shortcoming of this idea is that the system is not closed - the isothermal layer radiates longwave energy upwards into space. Hence, its temperature will drop and the inversion will weaken. Also, over the Arctic Ocean in winter the longwave radiation emitted from the surface is primarily balanced by both the downward longwave radiation and a conductive heat flux through the ice. His concept was important, however, in stating that the surface and temperature maximum layer are strongly coupled through radiative processes.

Overland and Guest (1991) examined the problem of the winter temperature inversion over the Arctic Ocean using a one-dimensional atmospheric radiation model. A steady thermodynamic model for the skin temperature was employed balancing the surface longwave radiation deficit with a conductive heat flux through the sea ice and overlying snow cover. No lateral heat advection by the atmospheric circulation was permitted. The model was calibrated against observed clear-sky temperature profiles. The modeled downward longwave radiation to the surface was less than observed. It was concluded that the "missing" downward radiation could be accounted for by including the emissivity effects of "diamond dust" in the lower troposphere (Curry, 1983; Curry et al., 1990). A correction for an ice crystal layer with an emissivity of 0.21 provided the necessary downward longwave flux to match observations.

After calibration, the model was initialized with a linear temperature profile and allowed to evolve over 90 days. Initially, a realistic temperature inversion was observed to develop. However, the temperature maximum layer radiated strongly into space, and cooled rapidly. With less downwelling longwave radiation directed downward to the surface, the surface temperature declined, less slowly than the temperature maximum layer because of compensation by the conductive flux through the sea ice and snow cover. In accord with the dependency of longwave emission on the fourth power of temperature, as the temperature of the profile declined over time, the cooling rate declined. The end result was an essentially isothermal profile. Adding an extra sensible heat flux from leads could not make up for the radiative loss to-space.

Overland and Guest (1991) then repeated the experiment by including a steady lateral heat advection from lower latitudes using values estimated by Nakamura and Oort (1988). Eureka! With the inclusion of advection to offset radiative loss to space, the model temperature profile evolved to develop a steady inversion. In summary, these results point out that while the winter inversion structure can be considered to a first order in the context of a radiative equilibrium (coupling the skin temperature to the atmospheric temperature maximum) an inversion cannot be maintained in the absence of lateral advection. These results also point to the significance of "diamond dust" in maintaining the downwelling longwave flux.

Cloud cover strongly modifies the inversion structure. As discussed, clouds increase the downwelling flux. Serreze et al. (1992b) compared winter mean temperature profiles over the Arctic Ocean (using rawinsonde data from the NP program) for relatively clear (<50% cloud cover) and relatively cloudy conditions (>50% cloud cover). For relatively cloudy conditions, the mean temperature difference between the inversion base and top was 9.2 K, compared with 13.2 K for clear skies. Temperatures were higher in the cloudy cases throughout the atmospheric profile from the inversion base to the inversion top, but with the temperature difference largest near the surface. The SAT (based on the lowest temperature level in the soundings) under cloudy skies was 245 K, compared to 238 K for clear skies.

In explanation, some of the upwelling longwave radiation from the surface and from the atmospheric layers below the cloud base that would escape to space is absorbed by the cloud and reradiated downwards. If we assume that there is no difference in advection between the mean clear and cloudy cases, then the reduction in the longwave radiative cooling rate will result in a warmer atmospheric profile. As the difference in the effective emissivity between the surface and temperature maximum layer is reduced, radiative equilibrium requires that the difference in surface temperature and that of the temperature maximum is less than for clear skies with the shape of the profile adjusting accordingly.

The surface net radiation budget turns positive in spring in response to solar input, helping to break down the inversion structure. The seasonal increase in cloud cover assists in this process. Once the snow over the tundra melts, the strength and frequency of inversions decline as sensible heating becomes stronger. Over Arctic lands during summer, convection with consequent release of latent heat is not unusual. Over sea ice, summer melting of the surface fixes the upwelling longwave radiation flux. One result of this is that the surface cannot adjust to achieve quasi-radiative equilibrium. However, the fixed surface temperature (~0 °C), is often associated with shallow surface-based inversions (Busch et al., 1982).

+1 0

Post a comment